Image processing apparatus, image processing method, and recording medium

Abstract

An image processing apparatus is configured to process an image. The image processing apparatus includes: a calculator configured to calculate a parameter for each of a plurality of areas and for each of a plurality of models describing a plurality of body shapes, each of the plurality of models having the plurality of areas, and probabilities indicating to which of the plurality of models input images input for learning belong; and a comparator configured to compare an input inspection image with the plurality of models whose parameters are optimized based on the probabilities.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]The present application claims priority under 35 U.S.C. § 119 to Japanese Patent Application No. 2016-049533, filed Mar. 14, 2016. The contents of which are incorporated herein by reference in their entirety.BACKGROUND OF THE INVENTION[0002]1. Field of the Invention[0003]The present invention relates to an image processing apparatus, an image processing method, and a recording medium.[0004]2. Description of the Related Art[0005]It is known to utilise a machine learning algorithm to interpreting given data. Examples of what is determined by interpreting data include the state of data in view of the future development in a match of shogi, go, or the like, whether or not an object captured in an image is a person, and how the person is captured and what is a background in a captured scene. Machine learning is used in a wide range of applications including text identification, speech identification, image identification, and prediction of fut...

AI Technical Summary

Benefits of technology

This patent aims to improve the accuracy of image inspection.

Problems solved by technology

However, the conventional material discrimination and the conventional defect inspection using machine learning are disadvantageous in that, if a plurality of targets of the same type but in different conditions are contained in an image, accuracy of material discrimination or defect inspection will be insufficient.

For example, when defect detection is performed on an image containing a plurality of targets, e.g., screws that differ from each other in thread starting position, it is possible that, even though the difference in the thread starting position is not a defect, a difference between images caused by a difference in the thread starting position is undesirably determined as a defect.

If a threshold value for this determination is set to a value that permits the difference, tolerance for defects is also widened, which undesirably decreases defect detection sensitivity.

However, the technique described in Japanese Unexamined Patent Application Publication No. 2005-265661 is not suitable for inspection of targets such that a plurality of target are in an image and that thus vary widely.

Images